10 BULLETIN 1276, U. S. DEPARTMENT OF AGRICULTURE 
The odds are then computed from the probability tables for differ- 
ent values of Z and n (the number of samples). In determining the 
average difference of the mean the computation was carried to only 
one decimal place. For the value of Z the results were computed 
to three decimal places throughout this bulletin.’ 
A difference in yield between a given variety and the standard 
Kharkof, of +2 bushels per acre with odds of 10:1, for example, indi- 
cates that in these experiments the variety has an average acre yield 
2 bushels higher than Kharkof and, based upon the results obtained, 
the odds are 10 to 1 that the variety is a higher yielder than Kharkof 
at that station. 
Variations in the climatic conditions prevailing in both the Great 
Plains and Great Basin areas cause wide fluctuations in the yields 
of wheat from year to year at any station. These fluctuations often 
markedly affect the relative yields of two varieties from season to 
season and cause the odds or probabilities for the recurrence of the 
differences to be rather low. 
The Kharkof variety is used as a standard for all comparisons 
discussed in this bulletin. Kharkof has been grown and is one of 
the leading varieties in nearly all of these experiments. It is listed 
under several C. I. numbers, including Nos. 1442, 1583, 2193, and 
2208. All of the above lots were obtained from the same locality in 
the Kharkof Government of Russia but at different times. The 
lots of wheat appear to be identical except for C. I. No. 2193, which 
as now grown contains spikes having long beaks resembling 
Beloglina. 
In the tables the yields and other data for Kharkof, C. I. No. 
1442, are used as the standard. If this lot of Kharkof was not 
grown data are used from C. I. No. 1583. If neither was grown the 
data from C. I. Nos. 2193 or 2208, in the order named, were substi- 
tuted. 
VARIETAL EXPERIMENTS 
The results of varietal experiments with winter wheat at 15 dif- 
ferent experiment stations are presented here. At all of these sta- 
tions the experiments have been conducted by the Office of Cereal 
Investigations either cooperatively with the State agricultural ex- 
periment station or some other agency or independently. The 
source of the data is stated in connection with each station. 
While comparative yields have been the most important results 
obtained, other factors have been studied. Complete agronomic 
notes have been recorded on the varieties at most stations. Sam- 
ples of most of the varieties also have been milled and the flour baked 
in the laboratory of the Milling Investigations Section, Grain Diyi- 
sion, Bureau of Agricultural Economics of the United States De- 
partment of Agriculture. In addition to yield, therefore, data on 
the following important factors are summarized: Days from emer- 
gence to maturity, height of plant, percentage of stem-rust infection, 
test weight per bushel, percentage of winter survival, percentage 
of crude protein, yield of flour, and volume of loaf. 
3 The writers desire to acknowledge the assistance of John R. Hooker in tabulating the yield data and 
in determining the odds by the *‘Student’”’ method, 
